Notte
Build and deploy reliable browser agents at scale
1.5K followers
Build and deploy reliable browser agents at scale
1.5K followers
One platform to build and deploy enterprise-grade browser agents. Managed infrastructure (sessions, proxies, identities, vaults) via a single API. Hybrid architecture combines deterministic scripts with AI reasoning for production reliability.
This is the 3rd launch from Notte. View more
Anything API
Launched this week
Many websites don't have public APIs. Anything API fills that gap.
Turn any browser work into a production-ready API. Describe the task, and our agents build a custom function that calls the site directly. Ship a custom API endpoint you can deploy serverless, schedule on Cron, or call via API.
Tell Notte what you need. We ship the function endpoint.
Free
Launch Team / Built With
Product Hunt
Notte
@curiouskitty yes we do. We leverage the full notte ecosystem (browser profiles, vaults, etc.) to build the automation
Notte
@curiouskitty for auth flows that require 2FA, we recommend using browser profiles and either login manually to persist the cookies, or use an Agent to login on your behalf. If you put your MFA secret inside our vaults, the agent can login super easily
I just used this for my platform One Pager and I gotta say wow!
This is much more than a standard chatbot.
One Pager rapidly creates websites for outreach, but does not have a public API at the moment. I entered my website, then anything API created an account, logged in starting understanding my platform and created an API to build website using my data structure.
This is something I haven't seen before and really impressed me.
Curios if there is a way to output the endpoints created into a document for external use?
Also how to you ensure that people aren't building APIs for sites that aren't theirs?
Notte
@jake_friedberg Re security - we're trying to enforce as much as possible a fair use policy through our agent prompt and safety railguard. Eg - Anything will not let you automate LinkedIn work due to their terms of services
Notte
@jake_friedberg hey Jake! thanks for the nice feedback. Yes the endpoint that gets created is actually hosted as a Notte function on Notte (You get a pointer to the Notte console from Anything). You can get the curl command to invoke the function from anywhere directly from the Notte console and also get observability and co from there. hope this helps?
This is pretty cool. Btw when the agent generates the API endpoint, how deterministic is the execution layer? Is the AI only used during the build phase and the runtime stays fully scripted?
Notte
@lak7 The AI tries its best to create a fully scripted runtime, but can also create hybrids where an AI agent itself can be used during the execution of the function if needed
This solves a real pain — so many useful sites still have no API. Turning browser workflows into callable endpoints could save teams a lot of brittle scraping work. Curious how it handles sites that change their UI often.
Notte
@allinonetools_net once the endpoint is built, subsequent runs call the site directly rather than re-navigating the browser UI, so UI changes don’t affect it after the initial build
Notte
@allinonetools_net Handling function rebuilds on sites updates will be our next move ;)
Great launch—this is exactly the gap most GTM teams hit when they need reliable data from tools that don’t expose APIs. Quick follow-up from my earlier thread: if I need session persistence + proxy rotation for long-running enrichment jobs, can Anything API pin identity state per endpoint (so retries don’t break auth flows)?
Notte
@danielsinewe not 100% sure i got the question; but anything is compatible with all the Notte stack so you can run w/ proxies enabled, run with session profiles, etc.
got it — super helpful. The endpoint-level profile/session pinning is exactly what I was trying to confirm for long-running GTM enrichment flows. If you add first-class retry policies (idempotency keys + backoff) in the function config, this becomes even easier to run safely in production cron jobs.
Notte
@danielsinewe 100%
Notte
@chriswyatt2 good question - to generate the API you'd need to run the automation in a browser session first, handle the browser infra, captchas, proxies, auth etc. Anything API uses the Notte CLI to generate the browser session and ships it as an endpoint (that you can deploy serverless, schedule on cron, call via API)
Notte
@chriswyatt2 Yeah as Sam said, main diff is the hosted service. You can run your claude code generate API so many times locally but if you need to run it at scale or when you sleep you better have a managed infra that also handles stealth with your volume
rtrvr.ai
Cool stuff!
Couple of questions:
1. What happens when the site's API's/layout changes? Is this just going to silently fail with 0 results?
2. So after creating the API is it still going to launch a browser and so this just a playwright script getting executed or is your agent writing a script to hit the underlying API's of the site?
Notte
@arjun_chintapalli Thanks Arjun! 1) For now the function would just return an error when you'll invoke it, so would fail if there's a massive API/layout change but not silently. 2) The idea is to avoid spinning up a browser as much as possible - this is the case when our agent found a 100% API path, otherwise it creates a hybrid flow, in that case we still use our session browsers.
rtrvr.ai
@ogandreakiro I have done this kind of scraping by just manually looking at the network requests in Inspector, pasting requests into Claude and Claude just wrote and executed a script and gave me the data for free.
Would this be the comparable for this?
Notte
@arjun_chintapalli Anything does that in the 1st place, but then also (and that's the main value prop) lets you run your job (scraping or automation) at scale concurrently 24/7 - which ppl would struggle to do on their local machine, or with home rolled browser infra